Next scheduled job whiles transform running

[ChrisL]

Sorry if this has been covered elsewhere, I've had a reasonable search but haven't found this particular question..

Got a couple of jobs that take quite a while to rollup and transform, in fact, taking just over 24 hours to do so. I was watching today when the scheduled daily jobs were due to run again, but since the transform was still running at the time, the scheduled job (which is just a forward incremental and not set to rollup) was apparently skipped and rescheduled for the next run time, meaning I won't have got a backup point for today. The transform job actually then finished shortly afterwards so I have since manually started the next run so I've got a backup from today, but obviously I don't want to always run this manually!

Just curious if this is expected behaviour, and if so, are there any thoughts on how to work around it to keep the daily scheduled jobs running even if the transforms are still running?

Wouldn't it be possible for the next scheduled job to be queued until the transform has completed and then run as normal to create the next forward .vib file.

[Gostev]

This approach can result in undesired consequences with backups starting too late and snapshot commits going into the production hours, so the current approach is simply safer. Out of no backup, and production down caused by backups and following snapshot commits still going at 9am, most customers prefer no backup.

I generally recommend that customers revise their backup architecture (storage performance and backup mode) as soon as they get close to 24 hours transforms with daily backups. It is important to realize that the amount of data will keep growing, and transform times are only going to get worse and worse if you do not act immediately. Allowing the delayed job starts does not solve the main issue, but only lets you "forget" about it for the next few months. Before you end up with no daily backup for a different reason (transform taking 36-48 hours). And this can come from nowhere - just a few big infrastructure changes (or even some scheduled defrags on multiple VMs), and your today's 24 hours all of a sudden turns into 48 hours.